Race, Class and Income Inequlity

Race, Class, and Income Inequality
Author(s): Erik Olin Wright

Reviewed work(s):
Source: American Journal of Sociology, Vol. 83, No. 6 (May, 1978), pp. 1368-1397
Published by: The University of Chicago Press
Stable URL: http://www.jstor.org/stable/2778109 .
Accessed: 21/11/2012 20:07
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to
American Journal of Sociology.
http://www.jstor.org
This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PM
All use subject to JSTOR Terms and Conditions
Race, Class,
and Income Inequality'
Erik Olin Wright
University of Wisconsin-Madison, Center for Research on Politics and Society,
and Institute for Research on Poverty
The basic thesis of this paper is that class, defined within the Marxist
tradition as common position within the social relations of production,
mediates racial differences in income returns to education. That is,
class position is viewed as a determinant of the extent to which educa-
tion can be transformed into income, and thus it is hypothesized that
much of the commonly observed racial difference in returns to education
is a consequence of the distribution of racial groups into class categories.
The results of the study strongly confirm this perspective: the differ-
ences in returns to education between black and white males largely
disappear when the regression equations are run within class positions.
One of the most consistent findings of research on racial inequality is that
black males receive considerably lower income returns to education than
white males. Weiss (1970, p. 154) found that, within specific age groups,
black males received significantly lower returns to education than white
males, whether education was measured as years of schooling or as achieve-
ment level. Siegel (1965) found that, net of occupation and region of the
country, the difference in expected incomes of black and white males in-
creased monotonically with education: at less than elementary education,
blacks earned $700 less than whites (net of occupation and region) in 1960;
at the high school level this increased to $1,400, and at the college level, to
$3,800. Duncan (1969) has shown that even after controlling for family
background, number of siblings, and occupational status black males still
receive lower returns to education than white males.2
1
This research was supported by funds granted to the Institute for Research on Poverty
at the University of Wisconsin-Madison by the Department of Health, Education, and
Welfare, pursuant to the provisions of the Economic Opportunity Act of 1964. I would like
to express my gratitude to Arthur Stinchcombe, Barbara Heyns, Michael Reich, and Luca
Perrone for their comments and criticisms on various aspects of this work and to Aage
Sorenson
for his skepticism about an earlier draft of this paper. I would also like to thank
James N. Morgan of the Institute for Social Research for making the data for this study
available. The opinions expressed herein are the responsibility solely of the author.
2
To my knowledge, the only study which claims to present findings different from these
results is the research of Ross Stolzenberg (1973, 1975). Stolzenberg estimates a rather
complicated income-determination equation within 67 detailed occupational categories for
both black and white males. He then compares the partial derivatives of income with
respect to education for the equations and finds that in nearly one-half of the occupational
? 1978 by The University of Chicago. 0002-9602/78/8306-0003$02.35
1368 AJS Volume 83 Number 6
Race, Class, and Income
Inequality
None of these studies, however, has controlled for class position as under-
stood within the Marxist tradition, that is, the position of individuals within
the social relations of production. The underlying premise of a Marxist class
analysis is that, while the diverse dimensions of social inequality cannot be
reduced to class inequality, class relations nevertheless play a decisive role
in shaping other forms of inequality. In the study of income inequality, this
implies that class relations organize the structure of income inequality in
the sense that class position shapes the ways in which other causes influence
income. If this notion is correct, then an analysis of racial differences in
income that ignores positions within the social relations of production is
incomplete. More concretely, if it is true that the returns to education vary
substantially across class positions, and if it is true that black and white
males are distributed quite differently across class positions, then much of
the racial difference in returns to education could in fact be a consequence of
the class distribution of races. This paper will explore such a possibility.
OPERATIONALIZING CLASS
Before developing a series of explicit hypotheses about the interactions of
race and class in the income-determination process, it is necessary to discuss
briefly how the Marxist notion of class will be operationalized in this study.3
When non-Marxists use the term "class," it generally designates a group of
people who share common "life chances" or market positions (Weber 1968,
p. 927; Giddens 1973; Parkin 1971, pp. 18-23), common positions within
status hierarchies (Warner 1960; Parsons 1970, p. 24), or common positions
within authority or power structures (Dahrendorf 1959, p. 138; Lenski
categories the partial derivative is larger for blacks than for whites. Thus, he concludes,
"Earlier findings suggesting high within-occupation racial differences in wage returns to
schooling . . . were probably artifacts of the gross occupational classifications used. These
past findings appear to have been produced by the tendency of black men to be con-
centrated in the lowest-paying detailed occupation categories within the major occupa-
tional group in which they are employed" (1975, p. 314). The problem with this conclusion
is that Stolzenberg uses a natural logarithmic transformation of income, whereas Siegel
(1965) uses raw dollars. This means that Stolzenberg is estimating rates of returns to
education rather than absolute returns. The absolute returns to education within the de-
tailed occupational categories may still not have differed significantly between blacks and
whites, but Stolzenberg's results do not demonstrate this. I ran Stolzenberg's equation
using the data in the present study, calculated the partial derivatives for all blacks and all
whites, and discovered that the rates of return for all blacks were significantly greater than
for all whites (Stolzenberg does not report the results for all blacks and all whites). Stolzen-
berg's results thus indicate that these higher rates of return to education for black men as a
whole can also be found within about one-half of the detailed occupations held by black
men. His results do not indicate that the absolute returns for black and white men are the
same within detailed occupations.
I
For a more detailed discussion of this operationalization of class, see Wright (1976b, pp.
137-39). For an extended discussion of the concept of class within the Marxist tradition,
see Wright (1976a; 1976b, pp. 20-90) and Poulantzas (1975).
1369
American Journal of Sociology
1966, p. 95). In contrast to these usages, Marxists define classes primarily
in terms of common structural positions within the social organization of
production. In contemporary American society, this means defining classes
in terms of positions within capitalist social relations of production.
For the purposes of the present analysis, position within the social rela-
tions of production will be defined according to two basic criteria: whether
or not the individual owns his or her own means of production, and whether
or not the individual controls the labor power of others (i.e., supervises
people on the job). These two criteria generate four basic class positions, as
illustrated in figure 1. "Ownership of the means of production" is opera-
tionalized by the question, "Do you work for yourself or someone else?"
For self-employed individuals, "control of the labor power of others" is
operationalized in terms of having employees; for individuals who are not
self-employed, this criterion is operationalized by the question, "Do you
supervise the work of others, or tell other employees what work to do?"
One further distinction will be made within this basic class schema.
Clearly, some of the people who are placed in the manager/supervisor
category are nominal supervisors. This would be the case, for example,
for the head of a work team who serves as the conduit for orders from above
but who lacks any capacity whatsoever to invoke sanctions on his or her
fellow workers. Proper managers are thus distinguished from nominal
supervisors by the question, "Do you have any say in the pay and promotions
of your subordinates?"
It is important to be clear about the logic underlying these class categories.
They are not simply proxies for occupations. "Occupation" designates the
technical function performed by individuals within the labor process;
"class" designates the social relationship within which those technical
functions are performed. While, of course, different class positions include
different mixes of occupations, every broad category of occupations is
Ownership of the Means of Production
YES NO
Control YES Employers Managers/Supervisors
over the
labor power
of others NO Petty Workers
Bourgeoisie
FIG. 1.-Criteria for class position. In the upper left hand quadrant, the term "Em-
ployers" is used rather than "Capitalists," since in the present study most of the individuals
in this category employ fewer than 10 workers. See Wright (1976a, pp. 35-36) for a dis-
cussion of small employers as a contradictory class location between the capitalist class and
the petty bourgeoisie.
1370
represented within each class category (Wright 1976b, pp. 168-73; Wright
and Perrone 1977).
HYPOTHESES
The empirical investigation will revolve around six hypotheses.4
1. Managers as a whole will receive much higher returns to education than
workers.-This basic result has already been established in an earlier study
(Wright and Perrone 1975, 1977). The rationale underlying this hypothesis
is based on an analysis of the specific position of managers within capitalist
social relations of production. Specifically, this analysis suggests that within
the managerial category there will be a strong link between income and
hierarchical position on the one hand and hierarchical position and education
on the other. The high returns to education within the managerial category
are a result of this double link.
First, examine the link between hierarchy and income. The behavior of
all employees within a capitalist organization is controlled by a combination
of repressive sanctions and positive inducements. As one moves up the
managerial hierarchy, however, the balance shifts between these two modes
of control. While repressive controls may be effective in creating conformity
to explicit rules, they are not terribly well suited to generating responsible
and enthusiastic job performance. Because the managerial hierarchy is one
of increasing responsibility (and, in a limited way, increasing power as
well), there will be a tendency for the behavior of higher managers to be
controlled more exclusively through a structure of inducements. The result
is that managerial hierarchies will be characterized by a steep income
gradient attached to authority position (Tannenbaum et al. 1974, p. 107),
even when the education of managers is held constant (Wright 1976b, pp.
235-38).
Second, examine the relationship between educational credentials and
hierarchical position. In both the working class and the managerial category,
education is in part a determinant of the value of the labor power of the
individual (or what non-Marxist economists typically call "human capital").
It is therefore to be expected that both workers and managers would receive
a positive income return to their education. However, among managers,
educational credentials serve a second function. In addition to creating
genuine skills, education also serves as an institutional mechanism for
legitimating inequalities of power within capitalist organizations. In practice,
4
All of the hypotheses that follow center on the relationship of the working class and the
managerial category to racial differences in 'returns to education. Since such a small per-
centage of black males are either capitalists or petty bourgeois it is impossible, using the
sample available for this study, to examine systematically the interactions of these class
positions with race.
1371
this means that there will be a general tendency for people with lower
credentials not to be promoted above people with higher credentials, and
thus there will be a tendency for managerial hierarchies to be characterized
by rather steep educational credential gradients (Tannenbaum et al. 1974,
p. 112).
The combination of this steep education gradient and steep income
gradient associated with hierarchy means that the managerial category as a
whole will be characterized by an especially high return to education. That
is, in addition to the direct return to education resulting from increases in
the market value of labor power (skills), which both managers and workers
receive, managers receive an additional increment of income for education,
stemming from the link between the legitimation function of education
within hierarchies and the use of income as a control mechanism within
authority hierarchies. (For a more detailed discussion of this interpretation,
see Wright and Perrone 1977; Wright 1976b, pp. 105-10.)
2. Black males will be more concentrated in the working class than white
males.-While we will not explore the actual mechanisms by which individ-
uals are sorted into class positions, it is nevertheless predicted that one
outcome of this sorting process is that blacks will be more heavily concen-
trated in the working class than whites.
3. When class position is ignored, black males will receive lower returns to
education than white males.-This is the standard finding in sociological
studies of racial differences in returns to education. It will be formally
tested in order to show that the usual results hold for the data used in the
present study.
4. Within the working class, the returns to education for black and white
males will be much more similar than for all blacks and all whites.-If at least
part of the overall difference in returns to education for blacks and whites is
a consequence of the class distributions of the two racial categories, then it
would be expected that within the working class itself the returns should be
much more similar. While white workers may be in relatively privileged
working situations as compared with black workers, neither white nor black
workers occupy positions of authority (by definition) and, thus, neither
receive the legitimation increment of returns to education discussed in
hypothesis 1.
5. Within the supervisor category, the returns to education for black and
white males will be more similar than for all blacks and all whites.-The
argument here is essentially the same as that presented in hypothesis 4. To
the extent that the overall racial differences in returns to education are a
consequence of class distribution, within a single class position-in this
case, the very bottom level of the managerial hierarchy-the returns for
blacks and whites should be much more similar than for all blacks and whites.
6. Within the managerial category, black males will have lower returns to
1372
education than white males.-The argument in hypothesis 1 concerning the
high returns to education of managers hinges on the dual link between
education and hierarchical position and between income and hierarchical
position. If a particular category of managers is highly concentrated at the
bottom of the authority structure, then this education-hierarchy legitima-
tion mechanism will tend to be attenuated. Although no data are available
in the present study concerning the hierarchical distribution of race within
the manager category, white males can be expected to be much more evenly
distributed throughout the hierarchy than black males. If this is the case,
then within the managerial category the returns to education for black
males should be considerably smaller than returns for white males.
DATA
The data for this study come from the eighth wave of the Panel Study of
Income Dynamics (Institute for Social Research 1975) conducted by the
Institute for Social Research (ISR) at the University of Michigan. While
in the original year of the panel study the sample was a random sample of
5,000 households, by the eighth year of the study, because of successive
nonresponses, the sample was no longer genuinely random.5 While this
probably will not seriously affect the regression results, it 'certainly may
affect hypothesis 2, concerning class distribution of races. Two other data
sets, the 1969 Survey of Working Conditions and the 1973 Quality of
Employment Survey (both from ISR), will thus be used for hypothesis 2.6
Throughout the analysis, the samples will be restricted to active participants
in the labor force.
EQUATIONS
In order to test the hypotheses about class and race interactions with returns
to education, two regression equations will be estimated for each of the
race-class groups being compared
:7
i Two things need to be noted concerning nonrandomness of the sample in the Panel Study
sf Income Dynamics. First, whenever an individual left the original household in the study
(because of divorce, high school graduation, etc.), the "split-off" was also included in the
subsequent years of the panel. Thus, the sample is not particularly skewed on age dis-
tribution. Second, a fairly complex system of weights has been devised to correct, at least
partially, for nonrandomness in nonresponse. Thus, the regression results in the present
study are probably reasonably reliable in spite of the nonrandomness of the sample.
6
A discussion of these data sets can be found in Wright (1976b, pp. 132-35).
7 The significance of the slope differences between groups will be assessed using the con-
ventional dummy-variable interaction model (Kmenta 1971, pp. 419-23). This means
that, if we are comparing the education slopes for two groups, the t-test would be:
t = (B11 - B12)/(S2Bll + V2S2B12)/(V1 + v2)
1373
Income a + b1Education, (1)
and
Income a + b1Education + b2Occupational Status + b3Age
+ b4Seniority + b5Father's Status + b6Father's Education (2)
+ b7Parental Economic Condition + b8Annual Hours Worked.
While the present study will not investigate the more complex structural
equation model which underlies the second equation, this equation is never-
theless important in assessing the extent to which the class interactions
observed in equation (1) may be consequences of the class distributions of
the various control variables in equation (2). Thus, for example, if in
hypothesis 1 the greater returns to education for managers were entirely
due to the occupational status distribution across class categories, then,
when occupational status is included in the equation, the differential returns
to education between classes should be substantially reduced. All of the
control variables in equation (2) are to a greater or lesser extent either
causes or consequences of education, and all of them can also be plausibly
thought to vary with class position. By controlling for these variables in
equation (2), we will be able to see if the differences in education coefficients
between the various groups being compared can be considered direct conse-
quences of class and race.8
where B11 is the education coefficient for group 1, B12 is the education coefficient for group 2,
S2B11 and S2B12 are the standard errors of the coefficients for groups 1 and 2, respectively, and
vP and v2 are the degrees of freedom for groups 1 and 2, respectively. The denominator in
this t-test is the standard error of the education X class-dummy interaction term in the
usual dummy-variable interaction model. It can be computed equally well from the
separate regression equations for the two groups using the above formula.
8
A brief comment about the kind of information contained in these regression equations
might clarify the logic of eq. (2). These equations describe characteristics of positions in a
social structure. While the data are tagged onto individuals, the equations themselves do
not adequately describe individual income-attainment processes, but only the income-
determination process for certain structural locations. Since we have no data on how indi-
viduals are sorted into structural positions, and since education obviously may play an
important role in this sorting process, the education coefficients in the various equations
may say very little about individual returns to education. What they describe is the return
to education which characterizes the position "worker," "manager," "black worker," etc.
Now, positions within a social structure can be characterized along many dimensions.
Marxists assert that the critical dimension of position within social structures is class rela-
tions. Other social scientists have argued explicitly or implicitly that occupational status or
the social backgrounds selected into positions are the key dimensions. Equation (2) in
effect holds constant a variety of positional dimensions which have some claim to being
important determinants of the relationship between education and income. To the extent
that class and race positions differ on education coefficients even when all of these factors
are held constant, these coefficient differences can be considered direct, unmediated
consequences of the class-race positions themselves. (See Wright and Perrone 1977, p. 38,
for a further discussion of this issue.)
1374
VARIABLES
The variables to be included in the analysis are measured as follows.
1. Income is measured by total annual taxable income received by the
individual. In addition to wage and salary income, this variable includes
income from assets, interest, and other sources of unearned income. Both
regression equations were also estimated for three other income variables:
annual earnings, an estimate of "permanent income" (an average of income
over the previous seven years), and imputed hourly wage (wage and salary
income divided by total annual hours worked). In none of the comparisons
of racial differences in returns to education did the results differ significantly
for these alternative income variables (see Wright 1976b, pp. 328-39).
2. Education is operationalized by a quasi-credential scale in which:
0 = no schooling or illiterate,
1 = less than elementary school,
2 = elementary school,
3 = some high school,
4 = completed high school,
5 = high school plus some nonacademic training,
6 = some college,
7 = college degree, and
8 = graduate training.
3. Occupational Status is measured by the standard Duncan SEI scores.
These scores were available only for the seventh wave of the panel study
(1974), while the class position questions were asked only in the eighth
wave. This means, in effect, that we have the individual's occupational
status score at the beginning of the year in which income was earned and
his class position at the end of the year (i.e., the beginning of the following
year).
9
4. Age is included in the regression both as a rough control for cohort
effects and as a rough measure of years of experience in the labor market.
5. Seniority designates the number of years the individual has worked for
the same employer, or the number of years an individual has been self-
employed in the same business.
6. Father's Status is measured by the average SEI score for the father's
gross-census occupation category. While this is clearly a much weaker
variable than a status score based on the three-digit occupation classifica-
tion, it is the best available from the panel study data.
I
One other minor point concerning the status variable needs mentioning: about 6% of the
sample represents "split-offs" in the 1975 year of the survey, that is, those who left a
household after the 1974 survey and set up a new family unit. For these people a three-
digit occupation classification was not available, and thus in these cases the Duncan score
is based on the average SEI value for the gross-census occupation categories.
1375
7. Father's Education is measured by the same scale as respondent's
education.
8. Parental Economic Condition is a scale reflecting the respondent's
subjective perception of parents' economic status in which:
1 = parents generally poor,
2
=
parents generally about average, and
3 = parents generally well-off.
9. Annual Hours Worked is a product of the number of weeks worked in
the previous year and the average number of hours worked per week.
The means and standard deviations for each of these variables for each
of the race and class groups included in this study are given in table 1.
RESULTS
Hypothesis 1: Managers as a whole will receive much higher returns to educa-
tion than workers.-Table 2 indicates that, in the simple regression of income
on education, workers receive $851 for each increment in education while
managers/supervisors receive $1,689. When the various control variables in
equation (2) are added, the returns for workers are $655, while the returns
for managers/supervisors are $1,169. In both cases, the difference in returns
is significant at the .001 level. This hypothesis is thus strongly confirmed.
The theoretical rationale for this hypothesis centers on the relationships
among hierarchical position, education, and income. If this rationale is
correct, then it would be expected that the high returns to education among
managers/supervisors would disappear if we examined a single hierarchical
level within the managerial hierarchy. In a limited way, we can test this
proposition by examining separately the returns to education among mere
supervisors (positions which involve no say over pay and promotions) and
proper managers, since mere supervisors can be considered the bottommost
level of the hierarchy. As can be seen in table 2, mere supervisors differ
hardly at all from workers in returns to education (although they receive
more income at every level of education), whereas proper managers receive
much higher returns.
One obvious objection to these results is that the true relationship between
educational credentials and income might be curvilinear. Thus, what appear
to be slope differences between workers and managers might
in fact simply
be a consequence of the two linear regressions reflecting different parts of a
single, nonlinear credential-income function. Figure
2
presents the relation-
ship between the linear regressions
in table 2 and the mean incomes for each
educational level for workers, managers, and supervisors. Table 3 indicates
the R2 for equation (1) using the credential scale and using a series of dummy
variables to represent the individual levels of the scale. Both the visual
1376
~~ cn r I b O crz
00 cE ( O. ,O U)cr + )CrzwN cz
Cl) k , N N U
.N w
- U
I m x co -) x (O,
- o
m O m
4
0B
U ) CN U
..s...*
z ~ I
Cl X I ) \_sy NsN
0 I
z I
X~~~~~~~~~~
_ - C q t- m in cl in 00 'IC ooq oo
t
x C -t-4otbo
I I
?4' Cd (O., o ssNo
--
(=
o-
( O. o s + ) m (O. c- o x Cr O. U-)
~
Cd ..... ...........
ClO
z I
O *_ J I ot11m tf)\te. ZO % O @N
C Ad
I r sSe? oO oO 0 O?
H ol~ I ~j1 o
124 12)i
t~~~~~~~~~ CN cI 00 m C) 00 m
e
m oo m +o C
Z4 - 4) J 94
Cl)
uD~~~~~~~~~~~~~~~~~~~
0
4, _o
U)
c)I I" Ocroo oo usoo'O r> r oo
? Cd
. . .
.Nr O Co .On c r o r r
.
A
4
cd
-,1O m C' M', 6
!4
-,4
00 6) o0 -,1 O0 m m oo (= t- CD ,O oo
* 0
. . ~
~~~~~. . . . . . . . . .
! . :~~~~~~~~~~~~l . o0
0'O , C -~C
-
O-q 0 C) C
C
o
!~~~~~~~~
m ~~~c .I,- -l) o t
- - -; - - -; - - -
-
$n4 *n d * . . . ..
1377
~~2 .124 .~~ .- >2
.
.) 212
2.) 2.)
~~~~~~~c.
* 2.)
C
0 C~~~~~~ 02 ~~~~~~~~ ~ ~ s4
MzC
1377~~~~~~~~~~~~~~c
CN CN
(O" Ul) 00
cl
0
. . .
C-q r. 0
t- 00
C)
0 41 . . .
00 U')
0
cdu
0
cq
0
U
Cd
CN
W 4,
Cd cq oo
0, cd
cli
41
Cd
4,
00
Cdo
. . .
cn U)
Z
00
. . .
0
Cd
P4
z
.0
41
Cd
Cd4, . . . . . .
00
00 C14
m cl-i 00 cl-i cl-i
. . . . . .
. . . . . .
.
. .
-- t- Ul) cl-i a, m 00 m -It =
in t- a, 00 m o Ul) Ul)
00
00
P4
'Cod
00
a, 00
.4.3
0
00
0
u
..
. . ... . . .
cd
--j
), : w
4
0 Cd
C/)
:s cn cn > cn
C/)
w cn pq pq ce pq pq pq
H
Cd
o ce
cn
13 78
0
CN
Cd
cd
0 0 Cd
U
.4 En
>
W
w
WO
C) C) Cd
0
v CN
0
Cd
0
41
En
o
41
o 41
cd
W o
Cd
r.
a
Cd
u 0
E
w u
0
u
Cd CN
cod
. . .
CN 00 0
cd
.1
Cd
cd 0
0
0
4,
'Cd -o
>,, 41
In
Cd Cd
Cd
0 In
41
u
'In.
M,.
1
T4
0
Cd
0
Cd
In
4 W . . .
Cd
P4
cd
t-
oo C14
Cd
cc,
4 Cd
0
Cd
C14 00
00
rA
+-j - .
En
Cd
6
0
0
0
-&-,0
o
cd 0 Cd
0 00 >
I"
4-1
0 0
s.,0
ed
,;: =$ 0
cd q cd
C) 'WII-
4.1
Cd
0r4
En -M
..",
cd
1-4
Di
U
w 0 En
Iz -.- . w cd
0 cd to
cn .0
M 0
W cd
o 0
ce
0
0
1379
inspection of figure 2 and the closeness of the R2 using the dummy variables
and a single scale in table 3 indicate that the relationship of credentials to
income is reasonably linear within each class category. The differences in
slopes thus cannot be interpreted as artifacts of a single curvilinear relation-
ship between income and credentials for both classes.
Hypothesis 2: Black males will be more concentrated in the working class than
white males.-Table 4 gives the class distribution of black and white males
based on an average of the 1969 Survey of Working Conditions and the 1973
$26,000
-
LINEAR REGRESSION
MEAN INCOME FOR EACH
$24,000
-
EDUCATION LEVEL /
, MANAGERS
$22,000 /
$20,000
-
$18,000
-
o $16,000
o
Z / SUPERVISORS
_
$14,000 //
WORKERS
< $12,000
-j
z $10,000
z
z
$8,000
$6,000 ,'
$4,000
$2,000
l l l I . l l l
0 1 2 3 4 5 6 7 8
0 0
E
a) 0) a
-C
co-
c E E E Efc
,, s
Ic E
o o 0a a' a- 2 0 0 0 0
Z 0t) LL L c/)I I I Z (10 ) O (JH
EDUCATIONAL CREDENTIALS
FIG. 2.-Relationship of income to education for workers, supervisors, and managers
1380
TABLE 3
COMPARISON OF EXPLAINED VARIANCE USING
EDUCATION SCALE AND EDUCATION DUMMIES
(Annual Income = Dependent Variable)
R2 Using
Education R2 Using
Credential Education
Scale Dummies
Equation for:
Workers ........
........
.066 .069
Managers/supervisors .. .135 .154
Supervisors ............ .065 .065
Managers
. .155 .180
Quality of Employment Survey and the distribution for the panel study.
Because of the problems of nonrandomness in the panel study, the average
of the two earlier surveys is undoubtedly a more accurate estimate of the
actual class distribution of races. It is clear from these results that black
males are indeed more concentrated in the working class than white males:
61% of all black males compared with only 40% of white males fall within
the working class.'0
Hypothesis 3: When class position is ignored, black males will receive lower
returns to education than white males.-Table 5 presents the regression
equations for all blacks and all whites, for black and white workers, for black
and white supervisors, and for black and white managers. Figure 3 graphi-
cally presents the results for the simple regression of income on education.
Table 6 presents the statistical tests for significance of the differences in
returns to education for various race-class comparisons. In the simple
regression of income on education, white males receive $1,419 for each
increment in education; black males receive only $860. When the controls
in equation (2) are added, the returns among white males decrease to $1,147
and among black males to $614. Both of these differences are significant at
the .001 level. Thus, as in most studies of black-white differences in returns
to education, black males as a whole do receive lower returns than white
males as a whole.
Hypothesis 4: Within the working class, the returns to education for black and
white males will be much more similar than for all blacks and all whites.-In the
simple regression of income on education, the returns for all black males are
less than one-half of the returns for all white males; within the working
10
If anything, these figures underestimate the proportion of the black male population in
the working class, since unemployed persons are excluded from both the Survey of Working
Conditions and the Quality of Employment Survey. If one assumes that most unemployed
black males belong in the working class, then the actual proportion of black males who are
workers would probably be closer to 70% or 75%, and of white males closer to 45% or 50%.
1381
OC) C14 1 t- V') I
C,4 t- I V') -4 1 : C
Cd
Cd
t-
Cd
oo "o -Cd
Cd
pq
t- in
Cd
Cd
Cd
t-
1-4 OC) 0
cn C14 OC) Ch
P4
0
cn
Cd
P"
u Cd
0
4-1
0
oo
tw
pq
Cd
P4
0
>4 Cd
ON
cn
0,
clq 00
cn
Cd
>4 Z
u
0
PI
>4 P4
0
cn
E
cn
Z H
-4
cd In C'4 in CN
0 En >4
z 0
0 4
P,
0
En
Cd
00
cd
NO 4.1
Cld
Cd
0
Cld
lu -0
4
0
En
W
41
4
Cd >
0
cld
Cd
U; En
0 rn
NO
En S-4 bD
S-4
S-4
-4 W
S-4
cd
0
> bID
Cd
4.,0
Cj
rn 0 +J
0 >
0
- S-4
1214 Cld
Pj
o
cn
1382
Z~~~~~~~~0 E
o o o
? d
I0mC
X
11 N .,, mS1. ... om .. , m0
0 11 ..
I
Y
cd 0
cnI : : : ? ': : : ??: : : ' I
Zd C' C N
oo o o
ms~~~~~
~
CN
o
o
inNT O
+% cO c 00 TS~~~~~~~~~~0
11 1
1~~~~~~~~~~~~~~~~~~~~o
C'
d
n Csn
) C
Z <, a I T t * * >t * * xo4 . .-
cS I
... . : : : : : ...
5
1
0l
*
0 ......................... .
I1 -q C1 0. d Y11 *N * Y o11 . .. Y.
C
a vN;XN;
00 00 ~1383
00
Cd
00 Cq
cn 0
mm t-
Cd 0 10
o
u
Cd
cn
cd
P,
d
cd
Cd
u 00 0 . . .
cd =1
00 00 lzt
Cd
r.
En
0
Cd
Cd
Cd
00
00
En
Cd 00
En
lt
0
00 C,4 "o
Cd
0
rn t- 0 t-
44
S-4
U-) rn >
w -4
S-4
> w
ga, 11
.. . . . .. . . .
cd, C W,-., II, Cd C cd cd
4., Cq 4- C4
-
-
cn cn
"n-, pq pq 11-S pq pq
6,Pq
PP &Pq , PP pq , PP
-4
Cd Cd
pq
pq
1384
It:$
C13
w
-M cnw
0
. . .
t- o,
O
:j
CN 00 "t
0 0
0 :t
. -0
. .
00 c
--.(
0 0
0
Cd
tn
Cd
0
0
tn ON O
Cd
00
Cd
. . .
cd
0 ci
C,)
--loo
0
cd
Cd
W
T, cd
tn
Cd
00
CN m 0
r. Cd
o 00
cn
cn
cn
0
00
Cd
C-d
r.
0
0 cn
Cd
06 Cd . . .
ul
o 75
tn
cn 0
bz
Wm
C"l m in
O m
Cd
0 00 00 Cd
tn
InWI
ce
00 U-)
O
CN
Cd
.45
od o
0
rn
rn
41
Cd
n
Cd aj od Cd
4-1
tn
C/)
I.-I pq 4)
pq IU-11 pq pq 12S pq 4 &Pq pq
Cd
4
pq
0
1385
ALL WHITES
$1 5,000
ALL BLACKS
Lu
0
z $10,000
-j
z
z
$5,000
I I I IIIII
0 1 2 3 4 5 6 7 8
EDUCATION WHITE MANAGERS
BLACK MANAGERS
L'
$1 5,000
-
WHITE SUPERVISOR
WHITE WORKERS
z ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~BAKSPRIO
_ / ~
~~-
-- BACK
WORKERVSO
$5,000
0 1 2 3 4 5 6 7 8
EDUCATION
FIG. 3.-Returns to education for blacks and whites in different class positions
1386
TABLE 6
COMPARISON OF RETURNS TO EDUCATION FOR BLACK
AND WHITE MALES ACROSS CLASS CATEGORIES AND
WITHIN CLASS CATEGORIES
RETURNS TO EDUCATION
Eq. (1) Eq. (2)
All black and white males:
Slope
difference ................ 959 533
t-value of difference ............. 8. 1* 3.6*
Black
slope
as %
white
slope.....
49 54
% difference in eq. (1) eliminated
by
the controls in
eq. (2) ... . ... 44
Workers:
Slope
difference ................ 192 7
t-value of difference ............. 1.8 <1
Black
slope
as
% white
slope.....
76 99
% difference in eq. (1) eliminated
by
the controls in
eq. (2) ...... ... 97
Supervisors:
Slope difference ........ ........ 93 -216
t-value of difference ............. <1I < 1
Managers:
Slope
difference ................ 938 1,991
t-value of difference ........... . 3.5* 6.3*
*
Significant at the .001 level (one-tailed test).
class, on the other hand, the returns for black males are slightly over 75%
of the returns for white males. What is more, when the controls in equation
(2) are added, the returns for black male workers become virtually identical
to the returns for all white workers, while the returns for all black males are
still only 54% of the returns for all white males. This suggests that, within
the wvorking class, most of the difference between black and white males in
returns to education observed in the simple regression of income on education
is mediated by the control variables in equation (2), whereas this is not the
case for all blacks and all whites.
Hypothesis 5: Within the
supervisor category, the returns to education for
black and white males will be more similar than for all blacks and all whites.-
As in the case of black and white workers, this expectation is strongly
supported by the results. In the simple regression, black male supervisors
receive only $93 less returns to education than do white male supervisors,
and in equation (2) they actually receive returns $200 greater (although the
difference is statistically insignificant).
Hypothesis 6: Within the managerial category, black males will have lower
returns to education than white males.-As predicted, in both regression
equations black male managers receive significantly lower returns to educa-
tion than white male managers. However, it was not expected that the
returns to education among black male managers would be essentially zero
1387
in the multiple regression equation. The expectation was merely that,
because of restrictions of blacks to lower levels of the authority hierarchy,
the hierarchical-promotion mechanism would be blunted among black
managers, and thus the returns to education would be less among black
than among white male managers. There was no a priori expectation that
those returns would disappear entirely in equation (2).
One possible clue to these results might be found in the occupational
distribution among white and black managers (table 7). As would be
expected, black male managers are considerably more concentrated than
white male managers among unskilled and semiskilled manual occupations
(38.5% compared with 13%, respectively). What is somewhat surprising is
the much higher proportion of black managers who are teachers compared
with white managers (13% compared with 2%, respectively). Expressed in
a different way, nearly 60% of the black managers in professional or techni-
cal occupations are teachers as compared with less than 10% of white
professional-technical managers. Remember, these are proper managers,
people who, unlike mere supervisors, state that they have some say in the
pay and promotions of their subordinates. This implies that teacher-
managers either occupy administrative positions within their educational
institutions or direct research projects in which they authorize the pay and
promotions of research staff (all but one of the teacher-managers were
college or university teachers).
If the regressions in table 5 are rerun excluding teachers from the man-
agerial category, the results conform much more to expectation (table 8).
Black male managers still have significantly lower returns to education than
do white male managers, but the returns are not nearly as small as returns
in the regressions that included teachers.
I cannot offer a particularly coherent explanation for why the presence of
TABLE 7
OCCUPATIONAL DISTRIBUTION AMONG MANAGERS
FOR BLACK AND WHITE MALES (%)
White Males Black Males
Professional, technical, and kindred ..... 24. 6 22. 5
Professionals .. ..................... 19.3 9.3
Technicians ........................ 3.0 0
Teachers ............ .............. 2.3 13.2
Managers and administrators ..... ..... 35.2 25.0
Sales ................................ 4.7 .8
Clerks ............................... 1.9 1.7
Craftsmen and kindred ....... ......... 20.5 11.5
Operatives, laborers, and miscellaneous.. 13. 1 38.5
Total ........................... o100 100
1388
TABLE 8
RETURNS TO EDUCATION FOR BLACKS AND WHITES
WITHIN THE MANAGER/SUPERVISOR CATE-
GORY, EXCLUDING TEACHERS
TOTAL ANNUAL INCOME
Eq. (1) Eq. (2)
Supervisors:
White males .732 761
(182) '208)
Black males .506 934
(217) (350)
Difference .226 -173
t-value .1.1 <1
Managers:
White males (N=385) .. 2,154 1,570
(256) (293)
Black males (N= 60)
. 1, 582 880
(476) (473)
Difference .572 690
t-value .2.0* 2.1*
NOTE.-Numbers in parentheses are standard errors.
*
Significant at .05 level.
so many teachers among black managers should have such a drastic effect
on the education coefficient in equation (2). Most likely, this result has
something to do with the interrelationship of education, occupational status,
and income among this specific subgroup of managers. To say this, however,
merely describes the problem; it does not provide a theoretical explanation.
DISCUSSION: THE INTERPLAY OF RACISM AND CLASS RELATIONS
It would be a mistake to interpret these results as indicating that all racial
discrimination is really disguised class oppression. While it is true that the
differential returns to education for blacks and whites largely disappear
when we control for class (except in the case of managers), this does not
imply that race is an insignificant dimension of inequality in American life.
The empirical and theoretical problem is to sort out the complex interplay
of racism and class relations, not to absorb the former into the latter.
The most obvious way in which racism intersects class relations is in the
social processes which distribute people into class positions in the first place.
In recent years, sociologists have devoted considerable attention to the
effects of racial discrimination on occupational mobility chances of blacks
as compared with whites. To my knowledge, there have been no studies
which systematically explore the role of racism in the distribution of individ-
uals into different positions within the social relations of production. Of
particular importance in such a study would be the social processes which
1389
select people into the managerial/supervisory category and the mechanisms
which regulate the promotion patterns in managerial hierarchies. Racism
would affect the distribution of races within authority structures in two
general ways. First, as in the sorting process for occupations, various forms
of racial discrimination affect access to the mechanisms which sort people
into the managerial hierarchy (educational credentials, connections, etc.).
Since, as was argued earlier, people with lower credentials will tend not to
be promoted above people with higher credentials, the result will be a higher
concentration of blacks at lower levels of the managerial structure and a
higher concentration of blacks in the working class. Second, and perhaps
more important, because of the necessity to legitimate the social relations
of domination embodied in managerial hierarchies, racism will tend directly
to prevent the promotion of blacks above whites. Of course, this does not
mean that blacks will never be promoted above whites. Particularly when
strong political struggles against racism occur, corporations and bureau-
cracies may see the imperatives of legitimation as requiring the acceptance
of some blacks into "token" positions of authority within managerial
structures. But in the absence of such struggles, it would be expected that
the logic of hierarchical domination within capitalist production relations
and the necessity of legitimating that domination would generate racist
patterns of recruitment into and promotion up managerial hierarchies.
The above argument about recruitment and promotion presupposes the
existence of racism. Given the presence of intense racist beliefs, it is easy to
explain why blacks will not be promoted above whites within hierarchies;
but this begs the question about the existence of racism in the first place.
It is beyond the scope of this paper to attempt a systematic account of the
origins of racism and the social processes which reproduce it in
contemporary
American society. What I will do is very briefly indicate the essential thrust
of a class analysis of the role of racism in American capitalism and
showv howv
the present study relates to that analysis.
A common mistake made by Marxists in analyzing racism is to assume
that all forms of racial discrimination are
unequivocally
functional for the
capitalist class. This is similar to analyses of the capitalist state which
argue
that every policy of the state is orchestrated
by
the
capitalist
class to serve
its interests. Such "instrumentalist" views of the state and
ideology minimize
the intensely contradictory character of
capitalist society.11 Capitalism
simultaneously undermines and
reproduces racism,
and it is essential to
disentangle these two tendencies if one is to
genuinely understand the
relationship between class and race in
contemporary capitalism.
One of the basic dynamics of capitalist development
stressed
by Marx,
as
11
For a critique of instrumentalist views in Marxist theory, see Gold, Lo, and Wright
(1975) and Esping-Anderson, Friedland, and Wright (1976).
1390
well as many non-Marxist theorists, is the tendency for capitalism to
transform all labor into the commodity labor power and to obliterate all
qualitative distinctions between different categories of. labor. From the point
of view of the capital accumulation process, the more labor becomes a pure
commodity, regulated by pure market principles unfettered by personal ties
and ascriptive barriers, the more rapidly can capitalism expand. In terms of
the logic of accumulation developed by Marx in Capital, there will therefore
be systemic tendencies in capitalism to reduce racial discrimination in the
labor market and to treat black labor power as identical with any other
labor power.
But this is only one side of the story. Capitalism is not just a system of
capital accumulation; it is also a class system in which workers struggle
against capitalists, both over their condition as sellers of labor power and
potentially over the existence of the capitalist system itself. Whereas the
essential dynamics of accumulation may lead to an undermining of racial
differences in the labor market, the dynamics of class struggle tend to
intensify racism. To the extent that the working class is divided along racial
and ethnic lines, the collective power of the working class is reduced, and
thus the capacity of workers to win demands against capital is decreased.
The result will be an increase in the rate of exploitation of both white and
black workers, although the effects may well be more intense for blacks and
other minorities than for whites.12 As a divide-and-conquer strategy, racism
thus serves the interests of capitalists, both as individuals and as a class.13
We thus have a basic contradiction: capitalism tends to undermine all
qualitative distinctions between categories of labor, but the capitalist class
needs those qualitative divisions for its own reproduction as the dominant
12
While Marxists have often claimed that racism hurts white as well as black workers,
until recently there have not been any systematic empirical investigations of this proposi-
tion. Two recent studies deal directly with this question. Reich (1971, 1973) shows that, in
the 50 largest SMSAs, the greater the racial inequality in median family earnings, the
greater the inequality of earnings among whites (as measured by the Gini coefficient for
earnings among whites) and the weaker the level of unionization. Szymanski (1976) shows
that, for the 50 states, the greater the inequality between black and white median
earnings,
the lower the median earnings of white males and the greater the inequality among whites
(again, measured by Gini coefficients). Both of these sets of results indicate that racism, in
dividing the working class, leads to an increase in exploitation of all workers.
13
The analysis of racism as a divide-and-conquer strategy has been perhaps the central
theme in Marxist treatments of the subject. Marx emphasized this issue in his various
discussions of the "Irish Question." In 1870, for example, he wrote, "The English bour-
geoisie has not only exploited the Irish poverty to keep down the working class in England
by forced immigration of poor Irishmen, but it has also divided the proletariat into two
hostile camps.... In all the big industrial centers in England there is profound antagonism
between the Irish proletariat and the English proletariat. The average English worker
hates the Irish worker as a competitor who lowers wages and the standard of life. . . . This
antagonism among the proletarians of England is artificially nourished and supported by
the bourgeoisie. It knows that this scission is the true secret of maintaining its power"
(Marx and Engels 1972, p. 162).
1391
class.14 Both forces operate. The actual balance between the two depends
upon a variety of historical factors. For example, under conditions of
extreme shortages of labor, obstacles to labor mobility in the labor market
are likely to be rather costly to individual capitalists, and thus it would be
expected that racial barriers would be more rapidly eroded by imperatives
of accumulation. On the other hand, when the supply of labor is relatively
abundant and when individual differences between laborers make little
difference to productivity (because of routinization, automation, etc.), those
strictly economic imperatives are likely to be weaker. The extent to which
racial or ethnic divisions within the working class are being deepened or
eroded in a given capitalist society cannot therefore be derived directly
from the abstract theory of capitalist economic development. It is only when
such abstract theory is linked to specific political and ideological develop-
ments that it becomes possible to assess the real dynamics of racism in a
given society.
While the present study does not deal with this historical process, the
data in the research nevertheless can be related to both sides of this con-
tradictory tendency within capitalism, that is, both to the perpetuation of
important racial divisions within the working class and to the common
situation of all workers as workers, regardless of race.
The data presented in figure 3 and table 5 clearly indicate that, while
black and white workers receive similar returns to education, black-worker
income is less than white-worker income at every level of education. One way
of assessing this gap in income is to see what the expected difference in in-
come betwveen a black and white male worker would be if they both had some
intermediate value on the independent variables included in the equation.
In table 9 this gap is calculated at levels of the independent variables halfway
between the means for each group in the comparison being made.15 As can
be seen from this table, the income gaps between races are large and sta-
tistically significant for both regression equations within each class category.
14
Both Marxist theory and neoclassical economics recognize that the inherent economic
logic of capitalism is to reduce progressively economic divisions between races, in terms of
both income and occupation. The difference (in these terms) between the two perspectives
is that neoclassical economics treats capitalism solely as an economic system and ignores
the fundamental class antagonisms within that system. Thus, the political and ideological
imperatives of controlling the working class play no role in the theory. Instead of seeing the
relationship of capitalism to racism as an intrinsically contradictory process, neoclassical
economics typically treats racism as a problem of individual "tastes" for discrimination on
the part of employers and workers (see Becker 1971, pp. 13-18).
15
Thus, in the simple regression of income on education for the comparison of black and
white workers, the income gap is assessed at a value of education equal to (Eblack worker +
Ewhite worker)/2.
The statistical significance of this income gap can be tested in a way exactly
analogous to the test of slope differences, only in this case a t-test is performed on the differ-
ences between constant terms adjusted to the appropriate values of the independent
variables (see Wright 1976b, pp. 55-57).
1392
Furthermore, the addition of the various controls in the multiple regression
equations reduces the total difference in mean incomes between races within
classes by no more than 50%, indicating that a substantial part of the
difference in mean incomes between races within classes should probably
be directly attributed to racial discrimination. In terms of Marxist theory,
these results strongly suggest that black workers are exploited at a higher
rate than white workers and that racism has generated real, material divi-
sions between races within the working class.'6
TABLE 9
INCOME GAPSa BETWEEN RACES WITHIN CLASS CATEGORIES
Mean Income
(.. .
.......$)
Eq. (1)b Eq. (2)b
All respondents:
Gap in income .............. ......... 5,308 3, 698 1,868
Black expected income as % of white ex-
pected
income ....................... 64 73 85
% difference in means eliminated by con-
trols ............................... ... 30 65
t-value of
gap ......................... ... 16.4* 8.1*
Workers:
Gap
in income
..
........... .. 2,870 2,203 1,428
pected income .75 80 86
trols ............................... ... 23 49
t-value of gap ............. ............ ... 11. 9* 8.9*
Supervisors:
Gap
in income .3, 872 2,896 2,140
pected income .69 76 82
trols ............................... . .. 25 45
t-value of gap ... .. . ... 8.6* 6.8*
Managers:
Gap in income .... ................. . 5,051 3, 707 3, 011
pected
income ....................... 72 79 83
trols .............................. ... 27 40
t-value of gap . ..................... ... 8. 0* 6.8*
a
Income gaps represent the difference in expected incomes for two groups evaluated at a level of the inde-
pendent variables in the regression equal to the average of their respective means on the independent variables.
b
Independent variables in eq. (1)
=
education only; independent variables in eq. (2)
=
education, age,
seniority, background, occupational status, and annual hours worked.
*
Significant at the .001 level (one-tailed test).
16
In order to interpret these results as indicating a higher rate of exploitation of black
labor, it is necessary to assume that two workers who have the same values on all of the
variables in eq. (2) will have essentially similar complexities of labor (i.e., embodied labor
in their own lab or power) and intensities of labor (i.e., pace of work within the labor
process). Since eq. (2) contains annual hours worked, if we accept the above two assump-
tions then two workers who have the same values on all the independent variables will
produce the same amount of total value in a year. Any difference in their incomes would
1393
Ultimately, the political thrust of the Marxist theory of racism hinges on
the other side of the contradiction: that, in spite of the divisive character
of racism and in spite of the material differences between black and white
workers which racism generates, workers of all races nevertheless share a
fundamental class situation and thus share fundamental class interests.'7
The central finding of this study-that black and white workers have very
similar returns to education-reflects this common class situation. Table 5
also indicates that black and white workers have very similar income returns
to occupational status (whereas whites taken as a whole have significantly
greater returns than blacks taken as a whole), which again reflects the
commonality of their class position.
Furthermore, it is easy to show that the income gaps between black and
white workers, while significant, are much smaller than the gaps between
either black or white workers and small employers. Figure 4 indicates the
expected incomes of each race and class category assessed at the level of the
independent variables of white employers.'8 The unadjusted mean income
of white workers is less than one-half that of white employers; the mean
income of black workers, on the other hand, is 75% of white workers' mean
income. In absolute dollar amounts, the mean white workers' income is over
$14,000 less than the mean white employers' income, whereas the mean
black workers' income is only $2,900 less than the mean white workers'
then reflect differences in the costs of reproduction of their labor power and thus differences
in rates of exploitation. Two factors could undermine this conclusion. First of all, the
reproductive costs of labor power are not represented simply by wages, but by fringe bene-
fits, state subsidies for education and other services, etc. To the extent that such additional
elements of the costs of reproducing labor power are themselves correlated with earnings,
we are, if anything, underestimating the differences in rates of exploitation by looking ex-
clusively at direct income. Second, if black workers as a whole are overqualified for the jobs
which they hold, wage differences could reflect in part social waste of potential surplus
value rather than superexploitation of black labor in the technical sense of the concept. If a
college graduate works on an assembly line, he or she produces no more surplus value and is
no more exploited than a high school dropout in the same production process. The com-
plexity of the college graduate's potential labor is being socially wasted in such a situation.
In the strategy used to compare black and white exploitation above, however, this situation
would appear as a greater rate of exploitation of the college graduate. There is no way, in
the present data, to differentiate between underemployment and more intense exploitation.
For a much fuller discussion of the relationship between econometric models of income de-
termination and the Marxist concept of exploitation, see Wright (1976b, pp. 120-31).
17
"Fundamental" class interests refer to interests defined across modes of production (i.e.,
interests in capitalism vs. interests in socialism), whereas "immediate" interests refer to
interests defined within a given mode of production. Black and white workers may well have
conflicting immediate interests under certain circumstances (as do many categories of
labor within the working class) and still share fundamental class interests. For a discussion
of the importance of the distinction between fundamental and immediate interests, see
Wright (1978, chap. 2).
18
This procedure is basically similar to the familiar cross-substitution technique employed
by Duncan (1969) and others.
1394
$ WHITE EMPLOYERS 100
%
$25,000 Z
$20,000 - t)-
LJ| WHITE MANAGERS
74%
< Z
2i
~~~~~~~~64
%
L
w<
()
$15,000
-BLACK MANAGERS 54%
0 2
-j WHITE WO RKERS
<
$10,000
-BLACK
WORKERS
LU
z LUJO
Z
F-
?
l ~ ~~~~~~~~~~
IL
<
$5,000 La
LU
U-
0
M EAN INCOME ADJUSTED INCOME ADJUSTED INCOME ADJUSTED
INCOME FOR EDUCATION* FOR EDUCATION FOR EDUCATION
AGE AGE
SENIORITY SENIORITY
STATUS STATUS
BACKGROUND BACKGROUND
HOURS WORKED
*
INCOME ADJUSTED TO WHITE EMPLOYERS MEANS
FIG. 4.-Comparison of income differences between races and classes
income. When the various controls in equation (2) are added, the expected
incomes of both black and white workers (evaluated at the white employers'
means on the independent variables) increase considerably. Yet, the differ-
ence between workers and employers is still considerably greater than the
differences between workers of different races. These results indicate that
compared with even small employers-let alone proper capitalists-the
common position of black and white workers within the social relations of
production generates a basic unity of economic situation.
The results of this study have important implications for both Marxist
and non-Marxist social scientists. For Marxists, they suggest the fruitfulness
of using quantitative techniques in exploring the interactions of class and
race, especially when social relations of production are explicitly included
in the research. Future work should attempt to examine longitudinal changes
in the interactions of class and race in the income-determination process in
order to specify more precisely the contradictory logic of the relationship
between capitalism and racial divisions discussed above. In such research,
it is particularly important to examine changes in racial inequality within
1395
the working class (rather than globally within the population as a whole) in
order to be able to assess the relative role of the accumulation and social
control imperatives in shaping patterns of racial inequality in contemporary
American capitalism.
For non-Marxist social scientists, the results of this study demonstrate
the importance of social relations of production in understanding social
inequality. At least in the study of racial differences in the income-determi-
nation process, the exclusion of class relations from the analysis leads to
basic distortions in the results. For many reasons, Marxist categories have
rarely been included in social surveys and virtually never in census surveys.
While the findings of this study hardly establish the overall validity of the
Marxist paradigm, they do indicate that any serious study of social inequal-
ity must attempt to measure social relations of production.
REFERENCES
Becker, Gary S. 1971. The Economics of Discrimination. Chicago: University of Chicago
Press.
Dahrendorf, Ralf. 1959. Class and Class
Conflict
in Industrial Society. Palo Alto, Calif.:
Stanford University Press.
Duncan, Otis Dudley. 1969. "Inheritance of Poverty or Inheritance of Race?" In Under-
standing Poverty, edited by Daniel P. Moynihan. New York: Basic.
Esping-Anderson, Gosta, Roger Friedland, and Erik Olin Wright. 1976. "Modes of Class
Struggle and the Capitalist State." Kapitalistate, nos. 4-5, pp. 186-220.
Giddens, Anthony. 1973. The Class Structure of the Advanced Societies. New York: Harper
& Row.
Gold, David, Clarence Y. H. Lo, and Erik Olin Wright. 1975. "Recent Developments
in
Marxist Theories of the State." Monthlly Review 27 (5): 29-42; 27 (6): 36-51.
Institute for Social Research. 1975. Panel Study of Income Dynamics: Procedures and Tape
Codes for 1975 Interview Year, Wave VIII. Ann Arbor, Mich.: Institute for Social Re-
search.
Kmenta, Jan. 1971. Elements of Econometrics. New York: Macmillan.
Lenski, Gerhard. 1966. Power and Privilege. New York: McGraw-Hill.
Marx, Karl, and Frederick Engels. 1972. Ireland and the Irish Question. New York: Interna-
tional Publishers.
Parkin, Frank. 1971. Class Inequality and Political Order. New York: Praeger.
Parsons, Talcott. 1970. "Equality and Inequality in Modern Society, or Social Stratifica-
tion Revisited." In Social Stratification:
Research and Theory for the 1970s, edited by
E. 0. Laumann. New York: Bobbs-Merrill.
Poulantzas, Nicos. 1975. Classes in Contemporary Capitalism. London: New Left.
Reich, Michael. 1971. "The Economics of Racism." In Problems in Political Economy,
edited by David M. Gordon. Lexington, Mass.: Heath.
. 1973. "Racial Discrimination and the White Income Distribution." Ph.D. dis-
sertation, Harvard University.
Siegel, Paul. 1965. "On the Costs of Being a Negro." Sociological Inquiry 35 (Winter):
41-57.
Stolzenberg, Ross M. 1973. "Occupational Differences in Wage Discrimination against
Black Men: The Structure of Racial Differences in Men's Wages, Returns to Schooling,
1960." Ph.D. dissertation, University of Michigan.
. 1975. "Education, Occupation, and Wage Differences between White and Black
Men." American Journal of Sociology 81 (September): 299-323.
1396
Szymanski, Albert. 1976. "Racial Discrimination and White Gain." American Sociological
Review 41 (3): 403-14.
Tannenbaum, Arnold S., Bodgon Kavcic, Menachem Rosner, Mino Vianello, and Georg
Wiesner. 1974. Hierarclhy in Organizations. San Francisco: Jossey-Bass.
Warner, W. Lloyd. 1960. Social Class in America. New York: Harper & Row.
Weber, Max. 1968. Economy and Society, edited by Guenther Roth. New York: Bed-
minster.
Weiss, Randall. 1970. "The Effect of Education on the Earnings of Blacks and Whites."
Review of Economics and Statistics 52 (May): 150-59.
Wright, Erik Olin. 1976a. "Class Boundaries in Advanced Capitalism." New Left Review
98 (July-August): 3-42.
. 1976b. "Class Structure and Income Inequality." Ph.D. dissertation, University of
California, Berkeley.
. 1978. Class, Crisis and the State. London: New Left.
Wright, Erik Olin, and Luca Perrone. 1975. "Classi sociale, scuola, occupazione e reddito in
U.S.A.: una analisi quantitativa sulle disequaglianze sociale in una societa postindus-
triale." Quaderni di sociolegia 24 (1-2): 57-91.
. 1977. "Marxist Class Categories and Income Inequality." American Sociological
Review 42 (1) : 32-55.
1397

Race, Class and Income Inequlity

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Race, Class and Income Inequlity

Uploaded by

Copyright:

Available Formats

Race, Class, and Income Inequality

Author(s): Erik Olin Wright

You might also like